Upcoming Mid-August Maintenance
On August 19th, several LinGA systems will temporarily taken offline as part of system repair, upgrade and transition to the Shared Computing Cluster (SCC). Details are as follows.
Current Planned Maintenance
LinGA Storage Upgrade
Following a major hardware failure in June, the distributed storage system on which most LinGA systems rely has been operating with several ongoing issues. While no data was lost, the underlying filesystem was damaged and the final repair requires a complete re-installation and upgrade of the system. We have spent the last month migrating and synchronizing data to safe locations in preparation for the upgrade
The re-installation will begin on August 19th, at which time the storage cluster will be placed in read-only mode for 48-hours for the final synchronization. The system will then be erased, upgraded and rebuilt and returned to service later in the week.
Note that this will prevent jobs on any cluster (RedStar, BlueIce, RubySky, SCC) from writing to the LinGA storage system. However, jobs will be able to read from the LinGA storage system and write to the SCC storage system.
RubySky Transition to SCC
As part of the transition to Boston University’s Shared Compute Cluster (SCC), RubySky will be relocated and integrated with SCC as part of a “Buy-In” process. The integration with SCC will eliminate RubySky as a standalone cluster and enable charge-free usage for these nodes for BUMC researchers on the SCC cluster (more information).
Jobs submitted after August 12th will have limited runtimes. On August 19th, any running or queued jobs on RubySky will be terminated. At this time, the entire RubySky cluster will be taken offline, upgraded and relocated to Holyoke, MA for integration with the Shared Compute Cluster (SCC).